Design and Implementation of Automatic Generation Algorithm for Advertising Artistic Design Based on Neural Networks

Abstract

With the growth of AI, especially breakthroughs in artificial intelligence technology, designers can use machine learning algorithms to generate advertising artistic designs automatically. Traditional grid search and random search methods require a large amount of computational resources and time costs, while particle swarm optimization (PSO) algorithms can find better combinations of hyperparameters in a shorter period of time. This article proposes a current neural network (RNN) algorithm based on PSO, combined with computer-aided design (CAD) tools, to be applied to the automation generation and optimization of advertising artistic design in order to improve design efficiency. By combining the RNN algorithm with CAD tools, designers can automatically generate advertising designs through the algorithm and then adjust and output them through CAD tools. The results show that the PSO-based RNN algorithm can generate more diverse shapes, colours, and textures, demonstrating high novelty and innovation. Moreover, the correlation and overall coordination between various design elements have also been significantly enhanced.